Issues with test data loader

[preetom-saha-arko]

How to evaluate the model on test set? When I run the command "python train.py --resume {saved_model_path} --dataset XXX --test", I am getting "Keyerror : 0". This error generates from the getitem function of test data loader.

Moreover, phrase_out and word_split are getting returned from the getitem function but not getting used in test_epoch function (located in train.py). Then what's the point of returning these two?

[preetom-saha-arko]

@nku-shengzheliu Can we have some chat regarding this issue? My email id: [email protected]